Hacker News Reader

If it didn't improve their revenues, they wouldn't be doing it.

I can assure you that this is absolutely not the case. If you look at Plotz's comment, it's exactly what most editors will say:

Pages that run too long can irritate readers. We run stories of 2,000, 4,000, even 6,000 words, and to run that much text down a single page can daunt and depress a reader. So pagination can make pages seem more welcoming, more chewable.

They really believe this.

It's important to realize that despite all the buzz about A/B testing in the industry, editors pretty universally believe they don't need to validate their opinions through testing. Why not? Because though school and their careers they've been told that their opinions matter, and, in fact, that's all their job really is: voicing opinions and making judgements, because they believe they have special insight into what readers want.

Look no further than Poltz, a writer and editor with no software design or development experience, who somehow believes he's qualified to make that statement.

I'll ignore your general disdain for writers and editors, and point out that not all readers want the same thing. I've pushed for non-paginated articles at my publication, but I admit that I'm not always happy with the result, and we do hear reader complaints about not being able to orient themselves in an article like this: http://spectrum.ieee.org/aerospace/astrophysics/the-cosmolog...

(hopefully a wider content-well after an upcoming redesign will help)

I'm curious to hear what people think of a hybrid pagination system, like that currently used by Computing Now magazine. I like that it gives the user a sense of the article length without interrupting the flow of reading. http://www.computer.org/portal/web/computingnow/careers/cont...

your general disdain for writers and editors

It's completely justified. For example: show me where in Poltz's long career as a writer and editor that he secretly squeezed in the years of UX, design, and development experience that qualifies him to make a statement about designing user interfaces.

The problem is that editors simply don't believe there is special expertise here or that it requires anything beyond their insight.

I admit that I'm not always happy with the result ... http://spectrum.ieee.org/aerospace/astrophysics/the-cosmolog...

This design doesn't incorporate many of the base standards for displaying long-form web content. Font size is way too small (should be 16-18px, 14px absolute minimum), column is too wide ("ideal" is 66 characters, though there's a wide range). A good starter article: http://informationarchitects.net/blog/100e2r/ There are many, many more.

Until that page is completely redesigned and returned to a readable baseline, it's difficult to discuss or test anything.

Agree completely with font size and column-width critiques, both of which are addressed in an upcoming redesign. And, for that matter, I've always had issue with Slate's UX and design choices.

>The fact is that editors simply don't believe there is special expertise here.

I agree that UX and design are often neglected when traditional publishers work in digital mediums. But I find it odd that you can appreciate the experience that leads to good UX and design but not the editorial and reporting experience that leads to good copy.

not the editorial and reporting experience that leads to good copy

Writing, content strategy, curation, copywriting, etc are obviously all important skills. However, being an editor does not remotely qualify someone to run a multimillion dollar business, design software, dictate engineering decisions, etc. Unfortunately, the industry is structured such that they are doing exactly that.

Is it any surprise that slate's division is operating at a loss? It's the norm in a broken industry.

Not to be too off-topic, but with a slight modification:

> However, being an [insert pre-digital mgr/executive]* does not remotely qualify someone to run a multimillion dollar business, design software, dictate engineering decisions, etc. Unfortunately, the industry is structured such that they are doing exactly that.

This is everywhere, I think. I've experienced first-hand the same problems in non-publishing industries where companies who rely on the web to do business are staffed with people who hardly understand the web making technical and software decisions, overriding software engineers and programmers making suggestions that are in opposition to the random thoughts in their heads. This is where the trump card of being higher up the ladder is played, and typically with detrimental results of varying degrees.

It's a serious problem in any business when web & technology decisions are still being made by people who don't understand the web & technology, and can't build it themselves.

Outside of the software industry, most businesses I've run across think being able to open a web browser or being in a mgmt/exec-level position qualifies them to actually make good software decisions.

* By pre-digital I mean old businesses that have moved to the web for various things, but are staffed by people who don't build software making key software decisions [e.g., electrical engineers, marketing execs, etc. making decisions on platforms, languages, and implementations while ignoring the advice of actual developers].

>A good starter article: http://informationarchitects.net/blog/100e2r/

Somewhat ironically, I find that page a little hard to read because the background is so bright. I find reading text on HN much easier with a darker grey background.

Re: hybrid pagination, I like the idea - but that particular implementation was a little wonky for me. Since the reading area was taller than my browser window, so I scrolled down to read the whole "page" and then had to scroll back up to continue reading. And since it seems to scroll based on scroll events from the browser, I wound up scrolling back to the first "page" in the article while trying to do that and the whole thing winds up being kind of a bummer.

I think a potentially nice implementation of this idea would be more like reading on a Kindle - text is broken into screen-sized chunks, and then there are low friction ways to go forward and backward "chunks".

Pagination can in fact increase reading time. Most readers never finish an article. In fact, most readers drop out after the second paragraph. Not because the articles are shit, but because people have a short attention span.

By adding a 'next page' button, you create an incentive to keep reading. By spiking the curiosity of the reader to find out what's behind the bend, they linger instead of leaving.

>there are copycats that don't do the testing, don't optimize for their niche of readers, and are probably losing both money and readers because of it.

I think this is much more common than sophisticated A/B testing.

At our publication, we were able to make the business case that the added pageviews from pagination didn't make much of an impact on the bottom-line. Because we generally run out of premium ad inventory each month anyway, most of the extra impressions were just for extremely low-revenue, Google-served remnant ads. Again, some publications thrive on volume, but it's a tough way to make money as CTR on Google ads is very low.

>If it didn't improve their revenues, they wouldn't be doing it.

The publishing industry has not always been about maximizing profits. The weird hybrid of advertising and reader-supported has created a different mentality in the business. There are plenty of things that publishers still do- like supporting foreign correspondents- that are driven more by the editorial side of than the business side.

Does this mean that your publication, you decided not to go for pagination?

That's right-it will be fully implemented in an upcoming redesign, but we've already shifted our production and design process to a non-paginated model. Usually when our articles were posted on Slashdot, Reddit, or here, readers used the single-page URL anyway. So the number of pageviews past page two was almost negligible.

Well, it's optimized to make the amount of money short term. This kind of testing doesn't tell you what the reader thinks of your site, which is kind-of the OP point.

The user might click through to all the pages, but obtain a general disdain for your site. Not saying that pagination enough to make users hate a site, but for me, personally, it doesn't exactly help.

I hate split articles too, but I believe this way of thinking is common mostly among tech-savvy people. People that don't know technology don't get annoyed by it or, if they do, just assume that this is how the technology works. Since online news are created for the vast majority of readers that are not part of our demographics, they will make more money by just splitting the articles anyway.

Good point.

If I'm simply tolerating some aspect of a site, it means that as soon as an alternate source of or means for accessing the for the same content shows up - I'll leave and never look back.

What's the difference between showing adverts on the next page or adverts further down the same page? They could continue all the way to the bottom of the page. Not that I want to see this but it would be an improvement on paged articles.

The comment higher up says A/B testing is used, that may be true for publishers that make use of performance ads. If you're a well known entity then you try to only accept brand ads (CPM - pay per impression/view), and those brand ads come with lots of conditions.

Advertisers want to appear near the top of the page or near the top of the content - that's why ad networks have rules about the placement of ads. Even Google (which is more performance than brand) limits the number of units that are displayed on a single page.

If your ads are at the top then there's a good chance that they will appear on the user's screen. If your ads are at the bottom and the user found the article boring then the browser was closed before seeing the ad, yet the impression was still paid for (hence, much lower rates).

This is the same with paged content, the user is equally less likely to see the adverts because maybe they don't view that page/scroll down that far.

I'm guessing the issue is that if you use paged content you know that the user has looked at that part of the article. If the adverts are all on the same page then an impression is not proof that a user has had that advert appear on the screen. Having said that if bootstrap's scrollspy can detect when you've scrolled content into view then I don't see why you shouldn't be able to track advert views in the same way.

Paged content gives site owners (most often, content farms) the ability to divide what would be a 10-second read into a 2-minute read by way of forcing users to keep clicking through. This can artificially juke pageview metrics, "engagement" metrics, bounce rates, etc., which means the site can a) sell more ads, b) command a higher premium on ads, or c) both.

Typically, it makes for a horrific reading experience. And I believe it will become bad for business, too, if enough high-quality content sites emerge and offer a far more compelling experience to draw readers away.

There's a hurdle preventing that from happening, however, and the hurdle is that users seem fairly site-agnostic in their content preferences. They consume information without a great deal of thought as to where it came from. And it's easier than ever to receive content now, through social channels, and never even have an inkling as to its original source or site. This liquidity and disposability of content means that content farms still have a big advantage over sites that attempt to make themselves into interesting, repeat-visit destinations.

Now, savvier and more intelligent readers -- the kind who probably frequent this community -- tend to avoid content farms, and probably don't click on articles entitled "10 Totally Weird Diet Tricks That Will Burn Away 10 Lbs in One Day!!!," requiring a 10-page click through to read. But sadly, a lot of people do click those things. A lot.

> What's the difference between showing adverts on the next page or adverts further down the same page?

If you are selling views, that will not wash with the advertiser as you can't reliably tell if the user has scrolled to see the lower adverts (whereas you can easily see if they clicked to see page two).

Of course even now your stats are going to over-count already (the ad may stil count as a view if there was an error on the ad server, or if the user had a proxy or plugin that blocked it (depending how that blocking is done), and so on) but that particular source of uncertainty is already accepted as an unavoidable consequence of the way the tech works, where the advertisers will see the "can't really tell if they saw the lower one" issue as a solved problem: just put the lower content and adverts on a new page.

You can use JS to tell how far the page has scrolled and therefore if an advert below the fold has been seen. They already do this on this particular site to only load images as they are needed.

Pageviews only matter in, obviously, a pageview driven ad model. In other instances, like the author mentions with Buzzfeed, pure pageviews aren't the entire goal. If publishers can get advertisers more interested in unique visitors and (especially) length of time on the site, then pageviews could be less important, and terrible practices like Mashable-esque page-reload sliders and six page articles can start to go away.

It's less pageviews and more reader engagement. You want to provide your salespeople solid evidence to convince the advertisers to pay the premium rate.

While perhaps the last part of the deal is measured in ad impressions, all the other facts and figures are important: time on site, pages per unique visitor, where the site's traffic is coming from, etc.

I agree with what you're saying about pageviews as a metric for reader engagement, but it still seems like in an ideal world if the advertisers had a solid empirical measure of ad effectiveness rather than just a count of the impressions, then pagination would have no impact on ad revenues.

the A/B test of one of my clients (in the b2c vertical) showed (for their site with their users) that people clicked more pages (which was kinda expected), and stayed longer on the website overall - with pagination (instead of endless scrolling).

Paginating lists or paginating arbitrarily within long text content? There is a massive difference between the two.

Wouldn't it be possible that the time lost paginating accounts for the longer stay?

some other metrics (we call them "item views") went up as well.

but it must be said: users of that site special, they actually want to find something, so they are (mostly) not casual users.

i remember a case a few years back on a travel site where we had a different experience, time on site went up after we ditched pagination (but that site also had horrible page load time, so every page view was a great opportunity to leave the site, so well, it's all interconnected)